South Africa has a high burden of drug-resistant tuberculosis (DRTB) and until recently, ototoxic aminoglycosides were predominant in treatment regimens. Community-based ototoxicity monitoring programmes (OMPs) have been implemented for early detection of hearing loss and increased patient access.
A longitudinal study was conducted to describe the service delivery characteristics of a community-based OMP for DRTB patients facilitated by CHWs as well as observed ototoxic hearing loss in this population.
A descriptive retrospective record review of longitudinal ototoxicity monitoring of 194 DRTB patients undergoing treatment at community-based clinics in the city of Cape Town between 2013 and 2017.
Follow-up rates between consecutive monitoring assessments reached as high as 80.6% for patients assessed by CHWs. Few patients (14.2% – 32.6%) were assessed with the regularity (≥ 6 assessments) and frequency required for effective ototoxicity monitoring, with assessments conducted, on average, every 53.4–64.3 days. Following DRTB treatment, 51.5% of patients presented with a significant ototoxic shift meeting one or more of the American Speech-Language-Hearing Association (ASHA) criteria. Deterioration in hearing thresholds was bilateral and most pronounced at high frequencies (4 kHz – 8 kHz). The presence of pre-existing hearing loss, human immunodeficiency virus co-infection and a history of noise exposure were significant predictors of ototoxicity in patients.
DRTB treatment with kanamycin resulted in significant deterioration of hearing longitudinally, predominantly at high frequencies. With ongoing training and supportive supervision, CHWs can facilitate community-based ototoxicity monitoring of DRTB patients. Current protocols and guidelines may require reassessment for appropriate community-based ototoxicity monitoring.
Tuberculosis (TB) is a communicable disease spread when people who are sick expel the TB-causing bacteria into the air (World Health Organisation [WHO],
Tuberculosis that is resistant to at least two of the most effective anti-TB drugs, rifampicin and isoniazid, is known as DRTB (Centers for Disease Control and Prevention [CDC],
Treatment of DRTB takes longer and requires drugs that are more expensive and more toxic than those used for the treatment of TB (WHO,
Concerns regarding the ototoxic nature of injectable antibiotics and the availability of novel, less toxic, more effective drugs led to an update of the South African Department of Health (DOH,
When the use of injectable ototoxic medications is unavoidable, audiological ototoxicity monitoring is essential to optimise hearing-related outcomes (WHO,
The South African Department of Health has committed to addressing the disparity in human resources for health by prioritising the integration of 50 000 community health workers (CHWs) into the PHC system by 2024 (DOH,
To improve the efficacy and efficiency for early detection of hearing changes, existing ototoxicity monitoring programmes (OMPs) and treatment effects should be evaluated so that ototoxicity monitoring guidelines can be adapted to specific settings (Dillard et al.,
This study was part of a larger, longitudinal descriptive retrospective record review of a decentralised community-based OMP for patients with DRTB facilitated by CHWs between 2013 and 2017. This specific OMP was selected for investigation as it offers a novel approach to ototoxicity monitoring for DRTB patients with a timeframe allowing for as many study participants as possible. The data were collected at community health centres and PHC clinics in two sub-districts of the city of Cape Town (CoCT), namely the [location masked for blind review] and [location masked for blind review] sub-districts. At the time of data collection, the sub-districts were characterised by a predominantly coloured (30% – 50%) and black African (19% – 46%) population who mostly resided in formal dwellings (68% – 87%) (CoCT,
This study included patients from a larger study who met the following selection criteria: patients who (1) received kanamycin, (2) were tested using conventional audiometry (0.25 kHz – 8 kHz), (3) had a baseline assessment conducted before, on the same day, or within 2 weeks of initiation of medication, and (4) had one or more follow-up monitoring assessments using conventional audiometry thereafter. The selection criteria were based on the OMP protocol and guidelines for ototoxicity monitoring (HPCSA,
Patient selection procedure.
The data collection procedure for this study was the same as for the larger study (Stevenson et al.,
Community health workers and PHC audiologists travelled to the clinics in each sub-district with portable audiological equipment. Community health workers and PHC audiologists were testers in the Mitchells Plain/Klipfontein sub-district, whereas only PHC audiologists were testers in the Western or Southern sub-district. At the time of a patient’s baseline assessment, identifying information including the patient’s name, date of birth, gender and medical history pertaining to HIV status, DRTB medication/s, comorbidities and adverse effects were recorded manually on a paper data collection form by CHWs and PHC audiologists. This information was obtained from the patient’s medical records in a clinic file and/or verbally reported to the CHWs and PHC audiologists during the patient interview. The KUDUwave portable audiometer (eMoyo, South Africa) was used by CHWs and PHC audiologists in this study. The KUDU wave is a PC (Dell laptop) controlled clinical diagnostic audiometer, and integrated supra-aural ear-cup and insert earphone headset, with an electronic response button for use without a soundproof booth. Automated and manual programmes conduct audiometry up to 16 kHz. Results are stored electronically and store-and-forward for printing.
The protocol for baseline and monitoring audiological ototoxicity monitoring assessments followed by the OMP at the time of data collection was as follows: a bilateral otoscopic examination was conducted followed by air-conduction pure-tone audiometry, and the findings recorded on the data collection form. If outer or middle ear pathology was suspected following otoscopy, the patient was referred to the managing doctor or nurse for appropriate treatment and referred for audiometry, according to the OMP protocol. Baseline assessments were conducted at the clinics prior to, on the same day or within 2 weeks of DRTB treatment initiation. Monitoring assessments were conducted once a month during the initial 6-month treatment regimen and then at 3, 6 and 18 months thereafter. Where an ototoxic shift meeting predetermined criteria (ASHA,
Each patient’s descriptive and audiological data were recorded manually by the CHWs and PHC audiologists on paper data collection forms and stored in the patient’s clinic file. A copy of each patient’s data collection form was kept with the CHWs and PHC audiologists and regularly made available for review to the managing PHC audiologist responsible for each sub-district. Upon completion of a patient’s DRTB treatment and ototoxicity monitoring, the form was stored permanently with the PHC audiologist responsible for each sub-district. The researchers collected the hardcopies of the patients’ data collection forms from the managing PHC audiologists in each sub-district for analysis, and these were returned upon completion of this study.
Data were imported from Excel into Statistical Package for Social Sciences (SPSS) software (version 27), after which descriptive statistics such as frequency distributions, measures of central tendency and measures of variability were used to present and interpret the data in a meaningful way. Data cleaning was performed where data erroneously captured by the CHWs, PHC audiologists and/or the researcher, such as dates, were corrected to be in a uniform format. In cases where data was accidentally not captured by the researcher, the data collection forms were re-examined to supplement any missing data. Because the data differed significantly from normality (Shapiro–Wilk
The OMP used paper data collection forms, which were manually completed by the CHWs and PHC audiologists for each patient. Where important data were missing, this was because data were not recorded on the data collection forms by the testers, and were therefore unavailable for inclusion in this retrospective study.
The study was conducted according to the guidelines of the Declaration of Helsinki and was approved by the Institutional Review Board (or Ethics Committee) of the University of Pretoria (GW20161128HS; 63/2017), the CoCT (7788) and the Western Cape Department of Health (WC_2017RP22_896). Owing to the retrospective nature of this study, consent to access the existing data collection forms on behalf of the patients was granted by the Western Cape Department of Health and the CoCT Health Department. All patient identifying information was kept confidential as patient records were given a numerical code in order to ensure anonymity during data collection and analysis. Data from the data collection forms were recorded on a password-protected Excel spreadsheet for later analysis by the researchers.
Of the 831 patients included in the parent study, 201 met the participant selection criteria and were eligible for inclusion in this study. Seven patients with results indicating technical or procedural issues related to their baseline and exit assessment audiograms (i.e. improved thresholds [> 50 dB HL] across all frequencies) were excluded. A final sample of 194 patients (
Participant description at the time of the baseline assessment (
Variables | % | |
---|---|---|
Not recorded | 33.0 | 64 |
Male | 35.6 | 69 |
Female | 31.4 | 61 |
DRTB and HIV co-infection | 24.7 | 48 |
Noise exposure | 20.6 | 40 |
Tinnitus | 18.0 | 35 |
Otalgia | 6.2 | 12 |
Hearing loss | 5.2 | 10 |
CHW | 76.3 | 148 |
PHC audiologist | 23.7 | 46 |
DRTB, drug-resistant tuberculosis; HIV, human immunodeficiency virus; CHW, community health worker; PHC, primary healthcare.
Community health workers tested 76.3% (148/194) of the patients in the study. There was a statistically significant difference (
Percentage of patients attending assessments following the baseline assessment according to tester.
The follow-up rates of the first six monitoring assessments for patients assessed by PHC audiologists (69.2% – 87.0%) were higher than for those assessed by CHWs (51.2% – 80.6%) (
Follow-up return rates and average days between consecutive pure tone audiometry assessments according to tester type.
Tester/Monitoring assessments | CHWs |
PHC audiologists |
||||||
---|---|---|---|---|---|---|---|---|
Follow-up rate (%) | Ave No. of days betweenassessments | SD | Follow-up rate (%) | Ave No. of days betweenassessments | SD | |||
1st – 2nd | 73.0 | 108/148 | 47.8 | 37.8 | 87.0 | 40/46 | 68.1 | 83.9 |
2nd – 3rd | 80.6 | 87/108 | 53.2 | 62.1 | 80.0 | 32/40 | 55.2 | 52.6 |
3rd – 4th | 74.7 | 65/87 | 63.4 | 75.4 | 81.3 | 26/32 | 48.4 | 45.4 |
4th – 5th | 63.1 | 41/65 | 52.3 | 44.3 | 69.2 | 18/26 | 87.8 | 107.3 |
5th – 6th | 51.2 | 21/41 | 49.7 | 42.1 | 83.3 | 15/18 | 65.7 | 37.8 |
SD, standard deviation; CHW, community health worker; PHC, primary health care; Ave no., average number.
More than half (51.5%; 100/194) of the patients presented with a pre-existing hearing loss at the time of the baseline assessment, where a hearing loss was defined as one or more hearing threshold > 25dB HL in one or both ears across all frequencies, increasing to 66.5% (129/194) at the time of the exit assessment. On average, a decline in hearing thresholds from the baseline to exit assessment was evident across all frequencies bilaterally, with the deterioration most pronounced at the high frequencies (
Mean hearing thresholds and deterioration of the left (a) and right (b) ears from baseline to exit assessment (
Mean baseline and exit assessment hearing threshold values and hearing deterioration for the left and right ears (
Frequency (Hz) | Mean baseline dB | SD | Mean exit dB | SD | Mean deterioration dB | SD | |||
---|---|---|---|---|---|---|---|---|---|
250 | 20.8 | 16.8 | 194 | 23.7 | 21.4 | 193 | −2.8 | 17.8 | 193 |
500 | 18.9 | 16.6 | 194 | 22.5 | 23.0 | 193 | −3.5 | 18.2 | 193 |
1000 | 19.3 | 16.2 | 194 | 23.7 | 24.3 | 194 | −4.3 | 20.1 | 194 |
2000 | 17.7 | 16.4 | 194 | 21.5 | 23.8 | 193 | −3.7 | 19.3 | 193 |
3000 | 21.3 | 16.9 | 20 | 34.1 | 29.4 | 27 | −9.1 | 22.2 | 17 |
4000 | 17.2 | 18.6 | 194 | 23.2 | 26.1 | 194 | −6.0 | 21.9 |
194 |
6000 | 21.7 | 20.1 | 170 | 29.6 | 27.0 | 179 | −8.0 | 22.2 |
168 |
8000 | 24.5 | 21.9 | 194 | 33.3 | 29.0 | 193 | −8.9 | 23.5 |
193 |
250 | 20.8 | 16.2 | 192 | 23.7 | 22.1 | 193 | −2.9 | 18.4 | 191 |
500 | 18.2 | 16.3 | 193 | 23.1 | 23.1 | 193 | −4.8 | 19.7 |
192 |
1000 | 17.0 | 16.1 | 193 | 21.5 | 23.6 | 194 | −4.4 | 19.5 |
193 |
2000 | 16.5 | 17.1 | 193 | 21.2 | 24.8 | 194 | −4.7 | 20.0 |
193 |
3000 | 19.0 | 18.9 | 20 | 28.8 | 27.1 | 25 | −7.1 | 21.3 | 17 |
4000 | 16.9 | 19.2 | 193 | 24.1 | 28.1 | 194 | −6.9 | 21.9 |
193 |
6000 | 20.4 | 21.1 | 169 | 29.6 | 28.6 | 182 | −8.9 | 21.8 |
167 |
8000 | 23.7 | 23.3 | 193 | 34.0 | 28.7 | 194 | −10.0 | 23.0 |
193 |
Hz, Hertz; dB, decibel; SD, standard deviation.
, statistical significance of
The patients’ hearing thresholds were compared according to various pure tone averages (PTAs) in
Mean baseline and exit assessment pure tone averages of the left and right ears (
Mean pure tone average values and hearing deterioration for the left and right ears (
PTA frequency range (Hz) | Mean baseline dB | SD | Mean exit dB | SD | Mean deterioration dB | SD | |||
---|---|---|---|---|---|---|---|---|---|
Left ear | |||||||||
Overall PTA (500–4000) | 18.3 | 15.6 | 194 | 22.7 | 22.5 | 194 | −4.4 | 17.6 | 194 |
LF PTA (250–500) | 19.8 | 16.3 | 194 | 23.1 | 21.8 | 193 | −3.2 | 17.2 | 193 |
MF PTA (1000–2000) | 18.5 | 15.7 | 194 | 22.5 | 23.3 | 194 | −4.0 | 18.6 | 194 |
HF PTA (3000–8000) | 20.7 | 18.6 | 194 | 28.2 | 25.8 | 194 | −7.6 | 20.6 |
194 |
Right ear | |||||||||
Overall PTA (500–4000) | 17.2 | 16.0 | 193 | 22.5 | 23.3 | 194 | −5.3 | 18.6 |
193 |
LF PTA (250–500) | 19.5 | 15.9 | 193 | 23.4 | 22.3 | 193 | −3.9 | 18.4 | 192 |
MF PTA (1000–2000) | 16.8 | 16.1 | 193 | 21.3 | 23.6 | 194 | −4.5 | 19.0 |
193 |
HF PTA (3000–8000) | 19.9 | 19.8 | 193 | 28.7 | 26.8 | 194 | −8.5 | 20.3 |
193 |
PTA, pure tone average; Hz, Hertz; LF PTA, low-frequency pure tone average; MF PTA, mid-frequency pure tone average; HF PTA, high-frequency pure tone average; dB, decibel; SD, standard deviation.
, statistical significance of
The presence of an ototoxic shift was determined according to the three criteria developed by ASHA (
Distribution of patients presenting with an ototoxic shift at the time of exit assessment.
ASHA ototoxic shift criteria | No ototoxic shift evident % | Group 1 |
Group 2 |
Group 3 |
||||
---|---|---|---|---|---|---|---|---|
Patients | 48.5 | 94 | 42.3 | 82 | 43.3 | 84 | 4.1 | 8 |
Left ear | 62.9 | 122 | 32.5 | 63 | 29.4 | 57 | 2.1 | 4 |
Right ear | 55.2 | 107 | 33.0 | 64 | 36.1 | 70 | 4.1 | 8 |
Bilateral (left and right) | 70.6 | 137 | 23.2 | 45 | 22.2 | 43 | 2.1 | 4 |
ASHA, American Speech-Language-Hearing Association.
Group 1 shift of ≥ 20 dB at a single frequency; Group 2, shift of ≥10 dB at 2 adjacent frequencies; Group 3, shift to ‘no response’ at three consecutive frequencies.
, 100/194 patients presented with an ototoxic shift that may have met one or more ASHA criteria: 16.0% (31) met one ASHA criterion, 33.0% (64) met two ASHA criteria and 2.6% (5) met three ASHA criteria.
The prevalence of hearing loss severity according to the revised WHO grades of hearing impairment is presented in
Prevalence of hearing loss severity for the left (
Category | Patients |
Left ear |
Right ear |
|||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|
Baseline |
Exit |
Baseline |
Exit |
Baseline |
Exit |
|||||||
% | % | % | % | % | % | |||||||
Normal hearing (-10 dB HL – 19.9 dB HL) | 77.8 | 151 | 74.2 | 144 | 68 | 132 | 60.3 | 117 | 72.7 | 141 | 66.5 | 127 |
Mild hearing loss (20.0 dB HL – 34.9 dB HL) | 19.1 | 37 | 15.5 | 30 | 23.7 | 46 | 23.2 | 45 | 17.5 | 34 | 16.5 | 32 |
Moderate hearing loss (35.0 dB HL – 49.9 dB HL) | 1.5 | 3 | 4.6 | 9 | 4.1 | 8 | 5.7 | 11 | 3.6 | 7 | 8.2 | 16 |
Moderately severe hearing loss (50.0 dB HL – 64.9 dB HL) | 1.0 | 2 | 2.1 | 4 | 1.5 | 3 | 5.2 | 10 | 2.1 | 4 | 1.5 | 3 |
Severe hearing loss (65.0 dB HL – 79.9 dB HL) | 0.5 | 1 | 1.0 | 2 | 1.0 | 2 | 0.5 | 1 | 1.6 | 3 | 2.6 | 5 |
Profound hearing loss (80.0 dB HL – 94.9 dB HL) | 0.0 | 0 | 1.0 | 2 | 0.5 | 1 | 2.1 | 4 | 2.1 | 4 | 2.6 | 5 |
Total hearing loss (≥ 95.0 dB HL) | 0.0 | 0 | 1.5 | 3 | 1.0 | 2 | 3.1 | 6 | 0.0 | 0 | 3.1 | 6 |
Unilateral hearing loss (< 20.0 dB HL in the better ear, ≥ 35. 0 dB HL in the worse ear) | 3.1 | 6 | 9.8 | 19 | - | - | - | - | - | - | - | - |
db HL, decibel hearing level.
, In the better ear.
, Pure tone average of 500, 1000, 2000 and 4000 Hz.
The presence of a pre-existing hearing loss at the time of the baseline assessment was a significant predictor of the deterioration of the overall PTA (0.5 kHz – 4 kHz) of the left (β = −8.750; 95% confidence interval [CI] [−14.953; −2.547];
The majority (76.3%) of patients in this study were assessed by CHWs, possibly because there were more CHWs (six) acting as testers than PHC audiologists (two). The follow-up rates between consecutive monitoring assessments for patients assessed by CHWs reached as high as 80.6%. In addition, the average number of days between assessments was lower for patients assessed by CHWs (53.4 days) than for those assessed by PHC audiologists (64.3 days). The follow-up rate of patients assessed by CHWs is better than the rate of a community-based DRTB treatment programme that included ototoxicity monitoring, where the loss to follow-up was reported as being high as 38% (Moyo et al.,
Numerous challenges to the implementation of ototoxicity monitoring exist, including a shortage of trained health care professionals and a lack of resources to conduct serial monitoring (Dillard et al.,
Although there were positive outcomes for community-based ototoxicity monitoring facilitated by CHWs, the OMP failed to meet some quality benchmarks pertaining to the frequency and timing of ototoxicity monitoring assessments, as stated in the guidelines (HPCSA,
Significant differences in the frequency of ototoxicity monitoring by CHWs and by PHC audiologists were identified in this study. The number of monitoring assessments attended by patients assessed by PHC audiologists (mean 4.3) was higher than for those assessed by CHWs (mean 3.3). In addition, almost a third (32.6%) of patients assessed by PHC audiologists attended the recommended six follow-up assessments, compared to 14.2% of patients assessed by CHWs. The reasons for patients assessed by PHC audiologists attending monitoring assessments with more frequency than those assessed by CHWs could not be established in this study. However, a possible reason may be the supervision and quality control provided by OMP managers of ototoxicity monitoring conducted by CHWs. For CHWs to fulfil their role successfully, regular training and supervision are required (WHO, 2007). Reports from sub-Saharan Africa indicate that the current provision of training for CHWs is not sufficient to improve the quality of care in this region (O’Donovan, O’Donovan, Kuhn, Sachs, & Winters,
The current OMP used paper data collection forms that were manually completed by CHWs and PHC audiologists. However, important demographic information, such as patient gender (33%), was not recorded by CHWs and PHC audiologists, and was, therefore, unavailable for analysis owing to the retrospective nature of this study. An effective OMP data management system enables the comparison of serial monitoring through reliable data recording; this is more efficiently achieved using an electronic data management system (Khoza-Shangase & Masondo,
In resource-limited countries such as South Africa, baseline audiometric assessments are often not conducted within the recommended timeframe, before ototoxic damage is likely to occur (Ganesan et al.,
A history of noise exposure was a significant predictor of hearing deterioration in the current study, with patients who reported previous exposure to noise presenting with 3.75 times the deterioration in hearing sensitivity compared with those with no history of noise exposure. A previous report concurred, indicating that patients with a history of noise exposure and aminoglycoside treatment had poorer high-frequency hearing thresholds than those exposed to noise without a history of aminoglycoside treatment (Khoza-Shangase,
The reported prevalence of ototoxicity varies widely and depends on various factors, including drug type and dosage, and patients’ demographic profile, such as age (> 60 years), the presence of mitochondrial mutations and exposure to loud noises (Ramma et al.,
Patients in this study presented with a bilateral decline in hearing thresholds in all PTA groups, with the most pronounced deterioration at high frequencies at the time of the exit assessment. Drug-resistant TB treatment using kanamycin therefore had a negative effect on the hearing status of the patients in this study, with clinically and statistically significant deterioration of hearing thresholds, most markedly in the high frequencies. The findings of the current study, therefore, support the implementation of OMPs for DRTB patients who are administered aminoglycosides, particularly as the latest WHO DRTB treatment guidelines (WHO,
The limitations of this study included the absence of quality indicators for audiometry conducted by CHWs and PHC audiologists. In addition, the prevalence of adverse side effects experienced by patients was not established by testers at the time of exit assessment. Immittance measures were not included as part of OMP protocol, and therefore, the prevalence of ototoxic hearing loss may have been influenced by the inclusion of patients presenting with middle-ear disorders. Important data pertaining to patient description and treatment were at times not recorded by testers and were, therefore, unavailable for inclusion in this retrospective study, and thus, may have caused research bias. Researcher and analysis triangulation were applied to reduce the effects of research bias. The use of a non-probability sampling method may limit the generalisability of the results of this study.
The findings of this study support the employment of CHWs to facilitate community-based ototoxicity monitoring of patients with DRTB. However, the findings reveal that over time, community-based OMPs for DRTB show gaps in service delivery practices, most notably in the frequency and timing of ototoxicity monitoring assessments. The possible reasons for this may highlight the need for ongoing training and supervision of CHWs using novel tools, such as smartphone technology and applications like WhatsApp. Drug-resistant TB treatment with kanamycin caused clinically and statistically significant deterioration of hearing thresholds in patients, most prominently at high frequencies. In this study, the patients co-infected with HIV, those with a pre-existing hearing loss and those exposed to excessive noise were at higher risk for developing ototoxicity-induced hearing deterioration. Patients presenting with these conditions should be identified and prioritised by OMPs for more vigilant ototoxicity monitoring and all-oral treatment regimens. South African OMPs need support and novel approaches for community-based ototoxicity monitoring, with revision of the current recommendations to best suit the South African context. These may include the widespread integration of ototoxicity monitoring services facilitated by CHWs into the existing decentralised, community-based PHC service delivery frameworks using a portable, automated technology with integrated data-sharing capabilities.
The authors acknowledge and thank the key role players who made this study possible through their support and participation.
The authors declare that they have no financial or personal relationships that may have inappropriately influenced them in writing this article.
L.J.S. and D.S. were responsible for conceptualisation and contributed towards the methodology. L.J.S., D.S. and M.A.G contributed towards the validation and visualisation. M.A.G conducted the formal analysis. L.J.S. was responsible for the investigation. Resources were provided by the Western Cape Department of Health and City of Cape Town Health Department.
L.J.S., D.S. and M.A.G. prepared and wrote the original draft and L.B.d.J assisted with reviewing and editing the manuscript. L.J.S. was responsible for project administration, while D.S, L.B.d.J. and M.A.G. provided supervision for the study. All authors have read and agreed to the published version of the manuscript.
This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.
The data that support the findings of this study are available on request from the corresponding author, L.J.S., for ethical reasons. The data are not publicly available as they contain information that could compromise the privacy of the research participants.
The views and opinions expressed in this article are those of the authors and do not necessarily reflect the official policy or position of any affiliated agency of the authors.